ClawHub
Back to skill

Security audit

Clawdwallet

Security checks across malware telemetry and agentic risk

Overview

This wallet skill is coherent with its stated purpose, but it gives an agent sensitive crypto-wallet authority with under-scoped signing controls and an unpinned external code install path.

Install only if you trust the publisher and can review or pin the referenced GitHub code yourself. Use a dedicated wallet with minimal funds, never import a primary seed phrase, require explicit confirmation for every signature or transaction, verify dApp origins and contract calls, and prefer a pinned commit or audited release before granting signing authority.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
95% confidence
Finding
The skill advertises broad, high-risk wallet control capabilities such as setting up agent-controlled wallets, connecting to dApps, signing transactions, and managing crypto across many chains without defining tight activation boundaries or approval constraints. In a wallet/signing context, ambiguous invocation conditions can cause the agent to engage in sensitive financial actions too readily, increasing the chance of unauthorized transaction signing, unsafe dApp interactions, or misuse by prompt injection from surrounding task context.

Missing User Warnings

High
Confidence
97% confidence
Finding
The documentation describes approving signature and transaction requests but does not prominently warn that blockchain signatures and sends can be irreversible, may transfer assets, grant token approvals, or authorize harmful contract interactions. In this skill's context, the omission is especially dangerous because the agent controls a multi-chain wallet and receives remote signing requests, so users or downstream agents may underestimate the consequences of approving a request.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.